【作者】 王鑫；

【导师】 曹宏美； 张健；

【作者基本信息】 内蒙古大学 ， 交通运输（专业学位）， 2025， 硕士

【摘要】 高速公路作为现代交通基础设施的核心组成部分,承担着多重关键作用,其运行稳定性直接关系到区域经济的发展和社会的日常运转。本文以内蒙古自治区高速公路为研究对象,基于复杂网络理论构建高速公路网络模型,并通过模拟仿真分析网络在出现内部局部故障或外部扰动情形下的脆弱性与韧性。首先,基于复杂网络理论,结合高速公路的空间布局和节点交汇特征,构建了更为符合实际情况的复杂网络模型。并从网络结构与空间特征的角度初步探究了内蒙古高速路网的整体特征和网络类型,为下一步路网的脆弱性与韧性分析奠定基础。在此基础上,综合度中心性、介数中心性和接近中心性,基于熵权法(Entropy Weight Method)与CRITIC法对指标进行赋权,构建融合加权的TOPSIS综合评价模型,即ECT(Entropy-CRITIC TOPSIS Method,ECT)综合评价法识别关键节点与关键边。考虑到路网存在的空间社区结构,在识别关键节点时,基于Louvain社区划分方法,从社区内部与跨社区连接两个方面对节点进行分类识别,以更全面细致地刻画节点的脆弱性。然后通过仿真模拟随机攻击与动、静态蓄意攻击下路网的多方面脆弱性,系统评估路网在不同破坏模式下的脆弱性演化过程。进而,对韧性曲线进行分析后,针对高速公路网络的恢复过程,选择随机恢复、基于重要性的优先恢复和动态最大增益贪婪恢复三种恢复策略,并通过仿真探讨不同恢复策略在恢复速度与恢复效果上的差异,比较得出具有最优恢复效果的策略。最后,进一步结合关键节点的地理空间分布特征,提出区域恢复优化建议。研究结果表明,基于ECT综合评价法识别出的重要节点和边对不同的网络攻击均表现出高度敏感性,且仿真结果表明跨社区桥接节点对网络整体性能的影响显著大于社区内部节点。在脆弱性演化模拟中,随机攻击下网络性能下降较为平缓,整体呈现较低的脆弱性;静态蓄意攻击在扰动初期对网络结构影响显著,但后期趋于缓和;而动态蓄意攻击则表现出最强的破坏性,在删除约30%的节点与边后,网络效率损失已达到80%左右,表明网络在面临持续且目标明确的攻击时,结构脆弱性迅速显现。在韧性恢复仿真中引入的三种恢复策略中,对比发现动态最大增益贪婪恢复策略在恢复速度和整体效果方面显著优于其他策略,进一步结合节点地理空间分布,识别出关键节点的空间集聚特征。研究所揭示出的高速公路网的结构脆弱性演化机制及其恢复响应特征,对于内蒙古高速公路网络的脆弱性评估、路网功能损失预警信号的捕捉,应对可能改变系统功能的扰动的快速响应及恢复策略选择具有理论指导作用。更多还原

【Abstract】 As a core component of modern transportation infrastructure,expressways play a pivotal role in ensuring regional economic development and the smooth operation of daily life.This study takes the expressway network of Inner Mongolia Autonomous Region as the research object.Based on complex network theory,a highway network model is constructed,and the vulnerability and resilience of the network under internal failures or external disturbances are analyzed through simulation experiments.First,a more realistic highway network model is developed by incorporating the spatial layout and node connectivity characteristics of the expressway system.The structural and spatial properties of the Inner Mongolia expressway network are preliminarily analyzed to lay the foundation for subsequent assessments of network vulnerability and resilience.On this basis,three centrality measures—degree,betweenness,and closeness—are integrated and weighted using the Entropy Weight Method and the CRITIC method.A composite evaluation framework based on the improved TOPSIS model,referred to as the ECT(Entropy-CRITIC TOPSIS)method,is proposed to identify critical nodes and edges.Considering the presence of spatial community structures within the network,the Louvain community detection algorithm is employed to classify nodes as either intra-community or inter-community,thereby enabling a more comprehensive assessment of node vulnerability.The vulnerability evolution of the network under random,static targeted,and dynamic targeted attacks is then simulated to systematically evaluate its response to different types of disruptions.Subsequently,based on the analysis of resilience curves,three recovery strategies—random recovery,importance-based static recovery,and dynamic greedy recovery—are implemented to simulate the network restoration process.The simulation results are compared in terms of recovery speed and effectiveness,leading to the identification of the most efficient strategy.Furthermore,by analyzing the spatial clustering patterns of critical nodes,region-specific recovery optimization suggestions are proposed.The results show that the critical nodes and edges identified through the ECT method exhibit high sensitivity to various attack strategies.Inter-community bridging nodes have a significantly greater impact on overall network performance compared to intra-community nodes.In vulnerability simulations,random attacks result in gradual performance degradation and reflect relatively low vulnerability;static targeted attacks cause significant early-stage disruption but stabilize over time;dynamic targeted attacks produce the most severe effects,with approximately70%of network efficiency lost upon the removal of 30%of nodes and edges,indicating rapid exposure of structural weaknesses.In the resilience recovery simulations,the dynamic greedy strategy outperforms the others in both speed and effectiveness of recovery.Finally,based on the spatial distribution of critical nodes,region-specific recovery recommendations are proposed.The findings of this study provide theoretical support and practical guidance for vulnerability assessment,early warning of network function loss,rapid response to targeted disruptions,and optimization of recovery strategies for the expressway network in Inner Mongolia.更多还原